OpenAI: gpt-oss-20b

openai/gpt-oss-20b

gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...

Prompt $/1M

$0.030

Completion $/1M

$0.130

Context

131,072 tok

Best provider

352 tok/s · Amazon Bedrock

Specs

Model id: openai/gpt-oss-20b
Display name: OpenAI: gpt-oss-20b
Modality: text->text
Tokenizer: GPT
Instruct type: —
Context length: 131,072
Max completion: 131,072
Prompt $/1M: $0.030
Completion $/1M: $0.130
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionno
Moderatedno

Benchmark scores

Each axis is one benchmark, normalized to the best model in the field.

No benchmark data for this model.

Speed by provider

Provider	tok/s	TTFT	Total latency
Amazon Bedrockfastest	352	811 ms	1,027 ms
Artificial Analysis	269	411 ms	—
Groq	172	683 ms	926 ms
WandB	162	265 ms	274 ms
NextBit	149	651 ms	987 ms
Clarifai	137	332 ms	482 ms
Google	134	530 ms	787 ms
Together	123	240 ms	366 ms
DeepInfra	107	311 ms	400 ms
Ambient	101	677 ms	1,293 ms
Parasail	96	414 ms	465 ms
Novita	75	393 ms	433 ms
Phala	57	618 ms	681 ms
Fireworks	56	411 ms	470 ms
Chutes	50	787 ms	787 ms
SiliconFlow	35	1,203 ms	1,346 ms
DekaLLM	15	993 ms	1,275 ms